That's what I don't get about a lot of medical testing.
I presume Adderall was approved based on tests that lasted a couple years at most. It definitely seems like there is a class of medicine that can be beneficial in the short-term but be likely to produce bad results over decades.
I assume opioids could fall into this category too. In the short-term they could make people happier and more productive, but also their is chance of developing a dependency that may not be noticeable in the data until much later.
If drug X was found to benefit 100% of people for the first 2 years of taking it, but 5% of people developed a dependency that landed them in rehab or worse after 5 years, would we want the FDA to approve it?
I have a lot of libertarian leanings so I am mostly fine with it, but it does seem like the boundaries for what we consider beneficial or not are pretty arbitrary.